INDEX

Explanations

mentions of the word "bs" along with varying numbers

oai_token-act-pair · gpt-3.5-turbo

the repeated occurrence of the term "bs" (presumably referring to a particular context or model)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 delinqu

-0.73

 JUSTICE

-0.67

 inequality

-0.66

 priority

-0.65

 Oriental

-0.63

 Equality

-0.63

 honorary

-0.63

 recognized

-0.62

 justice

-0.61

 honorable

-0.60

POSITIVE LOGITS

bs

4.48

BS

1.67

bers

1.63

bing

1.54

bians

1.53

bes

1.51

bsp

1.45

bies

1.42

bed

1.40

bsite

1.37

Activations Density 0.012%

mentions of the word "bs" along with varying numbers

the repeated occurrence of the term "bs" (presumably referring to a particular context or model)

No Comments

No Known Activations