INDEX

Explanations

statements revealing surprising or unexpected information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a realization or discovery that something is true

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cious

-0.91

shaw

-0.81

oulder

-0.78

resents

-0.78

ingham

-0.74

asus

-0.73

naissance

-0.72

ettlement

-0.70

aving

-0.69

avorite

-0.68

POSITIVE LOGITS

wagen

0.77

wards

0.74

 Technician

0.71

fitted

0.68

size

0.65

URA

0.65

lier

0.64

skirts

0.64

casts

0.64

rig

0.61

Activations Density 0.036%

statements revealing surprising or unexpected information

phrases indicating a realization or discovery that something is true

No Comments

No Known Activations