INDEX

Explanations

mentions of high levels/intensity/importance

oai_token-act-pair · gpt-3.5-turbo

references to high levels or high quantities of something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

tnc

-0.80

Advertisements

-0.69

eca

-0.67

 Barg

-0.66

aurus

-0.64

arium

-0.64

 Doct

-0.62

Dialogue

-0.62

 Rebell

-0.62

opsis

-0.61

POSITIVE LOGITS

lighting

0.90

 enough

0.85

enough

0.82

school

0.78

 levels

0.77

rising

0.77

rise

0.75

 bidder

0.75

 school

0.74

 stakes

0.73

Activations Density 0.043%

mentions of high levels/intensity/importance

references to high levels or high quantities of something

No Comments

No Known Activations