INDEX

Explanations

words related to communications or interactions between people, especially in the context of receiving or responding to messages

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

words that express uncertainty or lack of knowledge

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ortex

-0.74

itone

-0.68

 Handbook

-0.66

ffe

-0.65

irth

-0.64

essed

-0.64

ulators

-0.64

cffff

-0.62

veyard

-0.62

 horm

-0.61

POSITIVE LOGITS

 else

1.30

 substantive

1.10

 meaningful

1.06

 conclusive

1.05

 whatsoever

1.00

 unusual

0.99

 resembling

0.98

 worthwhile

0.96

 noteworthy

0.95

 nefarious

0.94

Activations Density 0.088%

words related to communications or interactions between people, especially in the context of receiving or responding to messages

words that express uncertainty or lack of knowledge

No Comments

No Known Activations