INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

roach

-0.86

aband

-0.84

istani

-0.82

asta

-0.82

oodle

-0.81

Asia

-0.79

mith

-0.79

hiba

-0.75

adesh

-0.75

inki

-0.75

POSITIVE LOGITS

dinand

1.04

 Ernst

0.78

 Griffith

0.74

Cly

0.73

 Claude

0.70

 Ferdinand

0.70

 Fleming

0.68

Cyr

0.67

 Maurice

0.67

 Petr

0.66

Activations Density 0.013%

names of individuals

proper nouns, particularly names of individuals

No Comments

No Known Activations