INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

references to political entities and institutions

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

 Helpful

-0.75

ricular

-0.70

worldly

-0.69

 Flavoring

-0.67

ornings

-0.65

 embodiments

-0.62

ESCO

-0.61

natureconservancy

-0.61

Ear

-0.60

 derog

-0.59

POSITIVE LOGITS

ariat

0.80

ptroller

0.76

icter

0.75

itol

0.74

 ambassador

0.71

 catalogue

0.70

itent

0.67

mbudsman

0.67

liner

0.66

 empire

0.66

Activations Density 0.379%

locations or places

references to political entities and institutions

No Comments

No Known Activations