INDEX

Explanations

sections of text indicating the speaker is reluctant to provide further details

oai_token-act-pair · gpt-3.5-turbo

phrases indicating reluctance to discuss certain topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

arov

-0.75

ãĥīãĥ©

-0.73

æ©

-0.70

 turnover

-0.65

ereo

-0.59

roots

-0.59

ushi

-0.57

ounter

-0.56

nell

-0.55

kefeller

-0.55

POSITIVE LOGITS

 haha

1.13

 lest

1.08

 ðŁĺ

1.03

 ðŁĻĤ

1.02

 here

1.01

;)

0.98

 rant

0.97

:)

0.97

:(

0.96

 suffice

0.93

Activations Density 0.312%

sections of text indicating the speaker is reluctant to provide further details

phrases indicating reluctance to discuss certain topics

No Comments

No Known Activations