INDEX

Explanations

reactions to events or situations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "reaction."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

²¾

-0.76

Prov

-0.71

idas

-0.68

owship

-0.66

ooth

-0.66

vor

-0.65

zu

-0.63

fitted

-0.62

uph

-0.62

 Teach

-0.61

POSITIVE LOGITS

 reaction

3.98

 reactions

3.36

 Reaction

2.90

 reacting

2.12

 response

1.98

 reacted

1.89

 react

1.79

 reacts

1.71

 responses

1.51

 backlash

1.48

Activations Density 0.028%

reactions to events or situations

instances of the word "reaction."

No Comments

No Known Activations