INDEX

Explanations

words related to injuries and violence

oai_token-act-pair · gpt-3.5-turbo

terms related to injury and casualties

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

perm

-0.73

 garn

-0.72

etics

-0.70

SpaceEngineers

-0.70

ramid

-0.70

gency

-0.69

cle

-0.68

orig

-0.68

etic

-0.67

 dictated

-0.66

POSITIVE LOGITS

 bystanders

0.82

jured

0.81

 Survivors

0.76

lehem

0.73

lyak

0.68

 responders

0.67

 survivors

0.67

adoes

0.67

merga

0.67

 Rico

0.64

Activations Density 0.035%

words related to injuries and violence

terms related to injury and casualties

No Comments

No Known Activations