INDEX

Explanations

the word "completely"

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing completeness or totality

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

pring

-0.77

 Springs

-0.68

maid

-0.68

osi

-0.66

rers

-0.66

hao

-0.63

 Pike

-0.62

actionDate

-0.62

RTX

-0.62

resso

-0.62

POSITIVE LOGITS

 unprepared

0.92

 disregard

0.92

 disreg

0.91

 oblivious

0.88

 devoid

0.88

 unaware

0.88

 unrelated

0.88

 ignored

0.88

 unsu

0.87

heartedly

0.84

Activations Density 0.058%

the word "completely"

phrases emphasizing completeness or totality

No Comments

No Known Activations