INDEX

Explanations

words related to negative actions or consequences

oai_token-act-pair · gpt-3.5-turbo

terms related to damaged or compromised reputations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

WER

-0.74

lication

-0.71

audi

-0.70

 tract

-0.66

family

-0.65

itative

-0.64

 Disorder

-0.64

planes

-0.64

itars

-0.60

Vs

-0.60

POSITIVE LOGITS

 tarn

1.34

red

1.05

wcsstore

0.95

 tainted

0.87

ned

0.86

ning

0.75

 tint

0.75

iances

0.73

edin

0.73

ipedia

0.70

Activations Density 0.017%

words related to negative actions or consequences

terms related to damaged or compromised reputations

No Comments

No Known Activations