INDEX

Explanations

words related to matching or compatibility

oai_token-act-pair · gpt-3.5-turbo

instances of the word "match" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

cember

-0.71

irtual

-0.69

raq

-0.68

pmwiki

-0.67

sie

-0.67

duc

-0.67

cknow

-0.66

ä¼

-0.66

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

-0.65

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

-0.65

POSITIVE LOGITS

 matches

0.80

ees

0.79

 match

0.77

 Match

0.75

ioned

0.74

xual

0.71

 matched

0.67

sticks

0.67

match

0.66

maker

0.66

Activations Density 0.019%

words related to matching or compatibility

instances of the word "match" and its variations

No Comments

No Known Activations