INDEX

Explanations

phrases related to negative consequences or burdens

oai_token-act-pair · gpt-3.5-turbo

mentions of "toll" indicating negative impacts or consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

RGB

-0.80

itals

-0.67

 Correspond

-0.65

 Pere

-0.65

ģĸ

-0.61

aples

-0.61

raised

-0.61

cius

-0.60

 Prix

-0.60

aunders

-0.59

POSITIVE LOGITS

 toll

1.20

 booths

0.96

 booth

0.86

 plaza

0.84

 Toll

0.79

 burden

0.78

staking

0.75

 Dickinson

0.73

 traged

0.72

atures

0.71

Activations Density 0.005%

phrases related to negative consequences or burdens

mentions of "toll" indicating negative impacts or consequences

No Comments

No Known Activations