INDEX

Explanations

words related to situations or events that emphasize a sense of continuation or persistence

oai_token-act-pair · gpt-3.5-turbo

references to evaluations or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

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

nob

-0.64

 Wise

-0.62

 expert

-0.61

 Cary

-0.61

kn

-0.58

 Podesta

-0.58

 magn

-0.57

 counsel

-0.56

 Chain

-0.56

Hub

-0.56

POSITIVE LOGITS

uation

4.44

uations

3.23

uated

3.10

uating

3.04

uates

2.83

uate

2.68

uity

1.61

uing

1.60

uality

1.47

uum

1.31

Activations Density 0.012%

words related to situations or events that emphasize a sense of continuation or persistence

references to evaluations or situations

No Comments

No Known Activations