INDEX

Explanations

instances of mentions of issues or problems

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate complex societal or ethical issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ournal

-0.80

ricanes

-0.79

thora

-0.78

gres

-0.77

undreds

-0.75

tsky

-0.75

uctions

-0.74

ousands

-0.74

ques

-0.74

aults

-0.73

POSITIVE LOGITS

 albeit

0.87

 insofar

0.83

 hence

0.79

albeit

0.78

 therefore

0.77

 whereby

0.77

 regardless

0.70

 although

0.69

 unlike

0.68

 wherein

0.68

Activations Density 0.688%

instances of mentions of issues or problems

phrases that indicate complex societal or ethical issues

No Comments

No Known Activations