INDEX

Explanations

expressions related to treatment or addressing issues

oai_token-act-pair · gpt-3.5-turbo

references to 'treat' and related concepts

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

BLM

-0.65

 Scotia

-0.64

 passing

-0.62

 account

-0.62

orah

-0.61

 Kickstarter

-0.60

OTA

-0.59

 Safari

-0.58

 recount

-0.58

 Pathfinder

-0.58

POSITIVE LOGITS

reat

1.26

resses

1.03

ress

0.95

hens

0.94

mosp

0.94

enium

0.94

htaking

0.93

idon

0.93

ressive

0.92

iful

0.91

Activations Density 0.008%

expressions related to treatment or addressing issues

references to 'treat' and related concepts

No Comments

No Known Activations