INDEX

Explanations

words related to measurement or quantity

oai_token-act-pair · gpt-3.5-turbo

references to measurement or evaluation scales

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

esson

-0.73

stitial

-0.71

 Gore

-0.71

luaj

-0.70

odore

-0.70

tek

-0.67

hiro

-0.66

vous

-0.66

åĳ

-0.65

draw

-0.65

POSITIVE LOGITS

 scale

1.28

 scales

1.03

 Scale

0.91

 invari

0.88

scale

0.79

itized

0.76

 replica

0.75

Scale

0.73

 gesture

0.72

 unbeliev

0.69

Activations Density 0.009%

words related to measurement or quantity

references to measurement or evaluation scales

No Comments

No Known Activations