INDEX

Explanations

statements that are described as clear or evident

oai_token-act-pair · gpt-3.5-turbo

instances of the word "clearly" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

umption

-0.75

alez

-0.72

oleon

-0.69

ucky

-0.69

uese

-0.68

anish

-0.67

urch

-0.66

ourke

-0.66

pring

-0.66

lÃ©

-0.64

POSITIVE LOGITS

 deline

0.94

 differentiated

0.82

 spelled

0.80

 distinguish

0.79

 identifiable

0.79

 marked

0.78

IFIED

0.76

Effective

0.76

 contrad

0.75

 differentiate

0.75

Activations Density 0.019%

statements that are described as clear or evident

instances of the word "clearly" in various contexts

No Comments

No Known Activations