INDEX

Explanations

words relating to steps or stages in a process

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Ste" or "Steele"

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

ABE

-0.74

uates

-0.74

 Parables

-0.65

CRIP

-0.65

ospace

-0.63

MX

-0.63

NCT

-0.62

PKK

-0.62

gdala

-0.61

é¾įåĸļå£«

-0.61

POSITIVE LOGITS

Ste

1.25

lla

1.00

rer

0.99

rers

0.90

vey

0.84

rence

0.84

uart

0.83

ste

0.82

fan

0.82

rett

0.82

Activations Density 0.003%

words relating to steps or stages in a process

occurrences of the name "Ste" or "Steele"

No Comments

No Known Activations