INDEX

Explanations

terms related to technical tasks or actions

oai_token-act-pair · gpt-3.5-turbo

terms related to actions and processes

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

hedral

-0.71

metics

-0.62

 regenerate

-0.62

arine

-0.62

addons

-0.61

 rusty

-0.61

 ILCS

-0.60

theless

-0.59

Cola

-0.59

ocre

-0.59

POSITIVE LOGITS

able

1.22

ability

1.18

ings

1.03

ables

0.96

athon

0.93

nings

0.87

outs

0.82

 initiation

0.80

ers

0.79

abilities

0.79

Activations Density 0.413%

terms related to technical tasks or actions

terms related to actions and processes

No Comments

No Known Activations