INDEX

Explanations

references to a specific individual named "Lindsey Graham"

oai_token-act-pair · gpt-3.5-turbo

mentions of Lindsey Graham and related entities in a political context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 collections

-0.82

 feats

-0.71

 warranties

-0.69

 autom

-0.68

 Assembly

-0.68

WWF

-0.67

ti

-0.65

 Cuomo

-0.65

errilla

-0.64

cci

-0.63

POSITIVE LOGITS

 Lindsey

1.95

acent

1.63

 Cors

1.53

BN

1.41

Lind

1.24

eye

1.06

named

1.03

side

0.96

 Alger

0.94

 Lindsay

0.91

Activations Density 0.037%

references to a specific individual named "Lindsey Graham"

mentions of Lindsey Graham and related entities in a political context

No Comments

No Known Activations