INDEX

Explanations

references to villains, often described in a negative context

oai_token-act-pair · gpt-3.5-turbo

terms related to villains in narratives

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

galitarian

-0.81

ollen

-0.79

sterdam

-0.76

independent

-0.76

ikk

-0.70

issance

-0.70

chedel

-0.69

choes

-0.69

apers

-0.68

bles

-0.67

POSITIVE LOGITS

ous

1.07

 villain

1.02

ously

0.93

 villains

0.92

 Bane

0.90

 mastermind

0.83

hattan

0.70

esses

0.70

 scourge

0.69

CVE

0.68

Activations Density 0.014%

references to villains, often described in a negative context

terms related to villains in narratives

No Comments

No Known Activations