INDEX

Explanations

proper nouns, specifically the name "AMY GOODMAN" with varying activations within the text

oai_token-act-pair · gpt-3.5-turbo

mentions of the person Amy Goodman

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

urger

-0.76

alli

-0.69

thro

-0.68

separ

-0.68

antic

-0.67

hered

-0.62

 revenge

-0.62

rating

-0.62

nu

-0.61

bowl

-0.61

POSITIVE LOGITS

 GOODMAN

2.05

ertodd

0.91

GROUND

0.84

 WATCHED

0.83

 COVER

0.83

IFIED

0.81

 EDITION

0.79

 WATCH

0.76

 PRESIDENT

0.76

ONSORED

0.75

Activations Density 0.015%

proper nouns, specifically the name "AMY GOODMAN" with varying activations within the text

mentions of the person Amy Goodman

No Comments

No Known Activations