INDEX

Explanations

specific names or terms associated with individuals or locations

oai_token-act-pair · gpt-3.5-turbo

proper nouns referring to individuals or specific entities

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

 dens

-0.67

 recomb

-0.67

 dawn

-0.64

 crim

-0.59

-0.58

 converge

-0.56

 conserv

-0.56

ymm

-0.54

 Prof

-0.54

 Migration

-0.54

POSITIVE LOGITS

's

0.73

 himself

0.73

orst

0.72

tsy

0.72

hyde

0.70

ouf

0.70

kowski

0.70

 accuser

0.69

neau

0.68

kson

0.67

Activations Density 0.271%

specific names or terms associated with individuals or locations

proper nouns referring to individuals or specific entities

No Comments

No Known Activations