INDEX

Explanations

phrases indicating a change or discontinuation of previous actions or habits

oai_token-act-pair · gpt-3.5-turbo

repeated expressions of cessation or changes in behavior

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

ortment

-0.61

 Proceed

-0.60

 utmost

-0.59

gged

-0.57

urat

-0.57

romy

-0.56

maximum

-0.54

 bounty

-0.53

 gist

-0.53

 Difference

-0.53

POSITIVE LOGITS

 anymore

0.91

 than

0.91

:(

0.84

nces

0.83

;)

0.76

adays

0.73

 unless

0.73

than

0.73

:)

0.72

Activations Density 0.030%

phrases indicating a change or discontinuation of previous actions or habits

repeated expressions of cessation or changes in behavior

No Comments

No Known Activations