INDEX

Explanations

words related to matching or compatibility

oai_token-act-pair · gpt-3.5-turbo

terms related to matching or compatibility

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hement

-0.82

################################

-0.74

uler

-0.73

raq

-0.72

cation

-0.68

irtual

-0.67

ijk

-0.65

stadt

-0.65

Discuss

-0.64

********************************

-0.64

POSITIVE LOGITS

ioned

0.83

sticks

0.78

maker

0.73

ees

0.71

tle

0.70

making

0.69

amorph

0.69

stick

0.67

 match

0.67

ingly

0.65

Activations Density 0.032%

words related to matching or compatibility

terms related to matching or compatibility

No Comments

No Known Activations