INDEX

Explanations

names of places or terms related to war

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "War."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

essee

-0.95

ħĭ

-0.84

Ħ¢

-0.79

etsk

-0.71

OGR

-0.71

 suspic

-0.66

İĭ

-0.66

 Incarnation

-0.66

aminer

-0.65

 drawer

-0.64

POSITIVE LOGITS

rior

1.45

riors

1.36

fare

1.30

lords

1.24

lord

1.18

locks

1.17

ped

1.04

horse

1.00

ping

1.00

lock

0.98

Activations Density 0.028%

names of places or terms related to war

repeated mentions of the word "War."

No Comments

No Known Activations