INDEX

Explanations

research-related terms

oai_token-act-pair · gpt-3.5-turbo

references to research institutions and studies

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

 disorderly

-0.76

sto

-0.72

footed

-0.71

 inclusive

-0.70

stood

-0.69

empt

-0.69

akespe

-0.68

pas

-0.67

 occasion

-0.67

outs

-0.66

POSITIVE LOGITS

 Research

3.57

Research

2.78

research

2.40

 research

2.20

 Researchers

1.84

 Studies

1.71

 researcher

1.60

Studies

1.53

Researchers

1.52

 Study

1.47

Activations Density 0.013%

research-related terms

references to research institutions and studies

No Comments

No Known Activations