INDEX

Explanations

phrases related to causes of negative events or situations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "caused" in various contexts

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

 Technique

-0.79

 motto

-0.73

 skelet

-0.65

 Fargo

-0.64

aeper

-0.64

halla

-0.62

 Brass

-0.62

 scrimmage

-0.61

 Niet

-0.60

iddler

-0.59

POSITIVE LOGITS

 havoc

0.96

 cele

0.86

uria

0.83

auga

0.81

netflix

0.80

parable

0.77

 irre

0.77

ãĥĨãĤ£

0.75

 irreversible

0.70

lling

0.69

Activations Density 0.025%

phrases related to causes of negative events or situations

instances of the word "caused" in various contexts

No Comments

No Known Activations