INDEX

Explanations

the string "OK"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "OK."

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

cycl

-0.74

requ

-0.73

dimension

-0.72

ministic

-0.66

lav

-0.65

 subsidized

-0.64

 vain

-0.63

 shortest

-0.62

lled

-0.61

ãĥĨãĤ£

-0.61

POSITIVE LOGITS

OK

1.34

lahoma

1.11

OK

0.93

AY

0.88

ettle

0.84

lihood

0.79

IER

0.76

 Okay

0.74

ELL

0.73

etheless

0.72

Activations Density 0.005%

the string "OK"

occurrences of the word "OK."

No Comments

No Known Activations