INDEX

Explanations

phrases related to entertainment and media

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "entertainment."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

issance

-0.94

cean

-0.77

zzo

-0.72

pton

-0.71

ichick

-0.69

phrine

-0.68

eden

-0.68

merce

-0.68

ivari

-0.68

cific

-0.65

POSITIVE LOGITS

RY

1.16

ENT

1.14

IFIED

1.09

LY

1.08

ENTS

1.04

IAL

1.02

URE

1.01

ITY

0.98

IUM

0.96

 CLIENT

0.93

Activations Density 0.007%

phrases related to entertainment and media

mentions of the word "entertainment."

No Comments

No Known Activations