INDEX

Explanations

words related to unexpected events or outcomes

oai_token-act-pair · gpt-3.5-turbo

references to surprises and unpredictability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ibur

-0.78

heid

-0.71

lite

-0.70

men

-0.70

glass

-0.66

halla

-0.66

ula

-0.66

craft

-0.66

ulk

-0.65

kees

-0.64

POSITIVE LOGITS

 unexpected

3.59

 unforeseen

2.50

 unexpectedly

2.12

 surprising

1.92

 unpredictable

1.85

 unintended

1.85

 unusual

1.65

 surprises

1.59

 improbable

1.59

 sudden

1.58

Activations Density 0.027%

words related to unexpected events or outcomes

references to surprises and unpredictability

No Comments

No Known Activations