INDEX

Explanations

specific acronyms, particularly UST, in the text

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "UST" and related contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ciplinary

-0.95

epad

-0.74

plings

-0.74

gression

-0.73

hered

-0.73

cum

-0.72

ãĥ©ãĥ³

-0.72

emonic

-0.71

hiba

-0.71

eworks

-0.71

POSITIVE LOGITS

EMBER

1.09

UST

1.07

RAL

1.06

ARD

1.03

MENT

0.96

OUND

0.93

IAN

0.89

LAND

0.88

AGE

0.88

IELD

0.84

Activations Density 0.032%

specific acronyms, particularly UST, in the text

references to the abbreviation "UST" and related contexts

No Comments

No Known Activations