INDEX

Explanations

the word "nothing" with a strong match

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "nothing."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

asio

-0.67

nikov

-0.66

ilt

-0.65

 Lanka

-0.65

ocard

-0.64

odan

-0.63

srf

-0.61

oak

-0.59

grad

-0.59

unci

-0.58

POSITIVE LOGITS

 else

1.44

Else

1.10

 whatsoever

1.04

 Else

0.90

 remotely

0.88

 resembling

0.85

ness

0.83

 except

0.81

except

0.80

nesses

0.79

Activations Density 0.051%

the word "nothing" with a strong match

occurrences of the term "nothing."

No Comments

No Known Activations