INDEX

Explanations

phrases related to actions or events that result in negative consequences

oai_token-act-pair · gpt-3.5-turbo

instances of the word 'causing' in contexts related to negative consequences

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

ymph

-0.77

ault

-0.72

town

-0.70

ael

-0.66

aeper

-0.64

 scrimmage

-0.64

 Phant

-0.61

ellen

-0.61

externalActionCode

-0.61

 Maple

-0.61

POSITIVE LOGITS

 cele

1.05

uria

0.79

 causing

0.78

 havoc

0.77

 exacerb

0.72

 sore

0.71

auga

0.70

 cancell

0.69

ĵĺ

0.68

 griev

0.68

Activations Density 0.010%

phrases related to actions or events that result in negative consequences

instances of the word 'causing' in contexts related to negative consequences

No Comments

No Known Activations