INDEX

Explanations

terms related to violent actions such as bombings

oai_token-act-pair · gpt-3.5-turbo

mentions of bombing in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

laus

-0.87

eva

-0.77

Lear

-0.77

PRE

-0.75

Perfect

-0.75

Vert

-0.72

ITY

-0.72

GRE

-0.72

learn

-0.71

Topics

-0.69

POSITIVE LOGITS

 bombing

1.40

 bombings

1.26

 raids

1.16

 bomber

1.08

 spree

1.04

 bombers

1.03

 barr

0.99

 bombard

0.97

 bombardment

0.93

 bombed

0.93

Activations Density 0.012%

terms related to violent actions such as bombings

mentions of bombing in various contexts

No Comments

No Known Activations