INDEX

Explanations

the word "nil" with high activation values

oai_token-act-pair · gpt-3.5-turbo

references to "nil" or similar terms indicating a lack of value or absence

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

RAW

-0.77

ebus

-0.75

hen

-0.74

Redd

-0.73

eeks

-0.72

heny

-0.71

Feed

-0.71

hetti

-0.71

versions

-0.70

Hop

-0.70

POSITIVE LOGITS

nil

1.46

nil

1.28

Nil

0.99

sson

0.83

una

0.81

ocent

0.78

 null

0.76

 NULL

0.74

NULL

0.74

 indemn

0.71

Activations Density 0.007%

the word "nil" with high activation values

references to "nil" or similar terms indicating a lack of value or absence

No Comments

No Known Activations