INDEX

Explanations

quantities, qualities, or states that are emphasized

oai_token-act-pair · gpt-3.5-turbo

phrases indicative of predictions, consequences, and significant changes

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

soever

-0.78

é¾įå¥ĳå£«

-0.76

ãĥ¼ãĤ¯

-0.64

 Rossi

-0.60

ĪĴ

-0.59

Pros

-0.59

Downloadha

-0.58

 Sites

-0.58

heres

-0.57

framework

-0.57

POSITIVE LOGITS

 downfall

0.87

 conclusion

0.80

 termination

0.79

 realization

0.77

 shootout

0.76

 deterioration

0.76

 eventual

0.74

 unintentional

0.74

 runoff

0.73

 unavoidable

0.73

Activations Density 0.317%

quantities, qualities, or states that are emphasized

phrases indicative of predictions, consequences, and significant changes

No Comments

No Known Activations