INDEX

Explanations

names of people detected in the text

oai_token-act-pair · gpt-3.5-turbo

proper names, specifically people's names

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

 Jagu

-0.81

 footballer

-0.80

âĶĢâĶĢ

-0.77

netflix

-0.72

 relegation

-0.71

ahime

-0.70

ultimate

-0.69

 realise

-0.69

 lockout

-0.67

 Cinderella

-0.66

POSITIVE LOGITS

 Wasserman

0.99

 Friedman

0.90

 Biden

0.90

 Bernstein

0.89

 Blumenthal

0.89

 Berman

0.89

 Stein

0.88

 Soros

0.88

 Pelosi

0.87

 Sachs

0.86

Activations Density 0.399%

names of people detected in the text

proper names, specifically people's names

No Comments

No Known Activations