INDEX

Explanations

pairs or groups of items that are being compared or categorized in contrast to each other

oai_token-act-pair · gpt-3.5-turbo

references to pairs or dualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ugu

-0.82

rir

-0.78

rolet

-0.73

lus

-0.71

 Caption

-0.67

lov

-0.66

icipated

-0.66

gow

-0.65

amaru

-0.65

Fed

-0.64

POSITIVE LOGITS

 thirds

1.43

 halves

1.10

 dozen

0.99

fold

0.97

 hundred

0.96

 thousand

0.88

 dimensional

0.85

 extremes

0.84

 tablespoons

0.83

 separate

0.81

Activations Density 0.120%

pairs or groups of items that are being compared or categorized in contrast to each other

references to pairs or dualities

No Comments

No Known Activations