INDEX

Explanations

arguments or statements made by individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "argue" and its variations, indicating discussions or debates

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

beam

-0.74

mental

-0.66

cler

-0.63

eco

-0.63

psy

-0.63

 vicinity

-0.61

 Radiation

-0.60

onder

-0.59

Es

-0.59

Bio

-0.59

POSITIVE LOGITS

 against

0.86

 arguments

0.86

 arguing

0.82

 convinc

0.82

uments

0.82

 forcefully

0.82

 vehemently

0.81

 persu

0.80

 passionately

0.80

 argue

0.79

Activations Density 0.023%

arguments or statements made by individuals

instances of the word "argue" and its variations, indicating discussions or debates

No Comments

No Known Activations