INDEX

Explanations

words related to processes of limitation or stoppage

oai_token-act-pair · gpt-3.5-turbo

variations of the word "st."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

merce

-0.94

EStream

-0.79

ļéĨĴ

-0.73

hower

-0.73

 GOODMAN

-0.70

ptive

-0.69

veyard

-0.69

thodox

-0.67

 chronological

-0.67

iferation

-0.66

POSITIVE LOGITS

uffed

1.20

amped

1.08

aves

1.08

roller

1.07

inging

1.04

itched

1.02

atted

1.02

ashing

1.01

rolled

0.98

ashed

0.98

Activations Density 0.012%

words related to processes of limitation or stoppage

variations of the word "st."

No Comments

No Known Activations