INDEX

Explanations

proper nouns related to specific geographical locations, particularly the Bronx

oai_token-act-pair · gpt-3.5-turbo

references to locations, particularly focusing on the Bronx and Harlem

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

OTE

-0.82

orthy

-0.81

ontent

-0.81

aking

-0.76

AKING

-0.76

akable

-0.76

lust

-0.76

soType

-0.75

neau

-0.74

OTT

-0.73

POSITIVE LOGITS

xual

1.24

Zoo

0.99

erick

0.80

cair

0.78

 Bronx

0.78

 Clown

0.70

plex

0.69

 boiler

0.67

eno

0.66

 Doodle

0.65

Activations Density 0.039%

proper nouns related to specific geographical locations, particularly the Bronx

references to locations, particularly focusing on the Bronx and Harlem

No Comments

No Known Activations