INDEX

Explanations

words related to physical features or attributes

oai_token-act-pair · gpt-3.5-turbo

mentions of "body" 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Kafka

-0.78

 Hoover

-0.70

 Dickens

-0.70

PBS

-0.68

 Fundamental

-0.65

Ans

-0.65

yrinth

-0.65

 Booth

-0.64

 Motorsport

-0.63

thouse

-0.61

POSITIVE LOGITS

guards

1.36

guard

1.16

building

0.98

builder

0.98

weight

0.97

anguage

0.94

builders

0.94

parts

0.91

bags

0.87

wash

0.86

Activations Density 0.023%

words related to physical features or attributes

mentions of "body" in various contexts

No Comments

No Known Activations