INDEX

Explanations

references to specific names, potentially related to people or places

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names related to individuals or characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Ŀ

-0.70

¤

-0.68

selves

-0.67

 Cortana

-0.64

 Lisbon

-0.64

ĩ

-0.63

tics

-0.62

pron

-0.62

ı

-0.61

 WATCHED

-0.61

POSITIVE LOGITS

bard

1.25

 Roth

1.13

roth

0.99

stein

0.98

haar

0.95

igan

0.89

punk

0.88

schild

0.88

heit

0.87

leigh

0.87

Activations Density 0.009%

references to specific names, potentially related to people or places

proper nouns, specifically names related to individuals or characters

No Comments

No Known Activations