INDEX

Explanations

questions starting with "how"

oai_token-act-pair · gpt-3.5-turbo

questions and expressions of uncertainty

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kus

-0.69

qt

-0.68

ipment

-0.67

ãĥ¼ãĥ«

-0.66

ixel

-0.65

table

-0.65

 rall

-0.64

thal

-0.63

phant

-0.63

iky

-0.62

POSITIVE LOGITS

 else

0.83

 anymore

0.74

 coincidence

0.73

 misunder

0.72

 anybody

0.71

 consolation

0.68

}}}

0.66

 bothered

0.66

 explan

0.64

 Loving

0.64

Activations Density 0.116%

questions starting with "how"

questions and expressions of uncertainty

No Comments

No Known Activations