INDEX

Explanations

phrases related to different parts or sections of a larger entity or content

oai_token-act-pair · gpt-3.5-turbo

mentions of different segments or parts of a project or discussion

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

swe

-0.73

 skulls

-0.66

 grasp

-0.65

pal

-0.65

cro

-0.65

 shooter

-0.65

 grave

-0.63

tle

-0.63

 peanuts

-0.63

 mustard

-0.63

POSITIVE LOGITS

Part

3.70

part

2.42

 Part

2.22

PART

1.90

Parts

1.80

 PART

1.73

 Parts

1.47

parts

1.47

 part

1.45

Volume

1.30

Activations Density 0.015%

phrases related to different parts or sections of a larger entity or content

mentions of different segments or parts of a project or discussion

No Comments

No Known Activations