INDEX

Explanations

numerical ranges expressed in a particular format

oai_token-act-pair · gpt-3.5-turbo

sections related to scores and statistics in sports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 detection

-0.68

clo

-0.62

 appro

-0.61

 appraisal

-0.60

BDS

-0.60

 Narc

-0.60

 literature

-0.60

 solicitation

-0.59

 puberty

-0.59

 solicit

-0.59

POSITIVE LOGITS

straight

1.11

year

1.03

iron

0.99

percent

0.97

degree

0.97

win

0.97

game

0.97

inch

0.96

plus

0.95

Activations Density 0.085%

numerical ranges expressed in a particular format

sections related to scores and statistics in sports

No Comments

No Known Activations