INDEX

Explanations

proper nouns or specific names, especially those with a title of "first" or "last."

oai_token-act-pair · gpt-3.5-turbo

references to first names and their use in various 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

rencies

-0.79

assies

-0.79

 Machines

-0.74

 embassies

-0.70

abis

-0.70

 hordes

-0.69

Dru

-0.66

agements

-0.66

tics

-0.66

 repositories

-0.65

POSITIVE LOGITS

 contender

0.80

stroke

0.72

word

0.72

gradient

0.70

poke

0.70

prising

0.67

 broom

0.67

 apiece

0.67

hander

0.66

gap

0.65

Activations Density 0.229%

proper nouns or specific names, especially those with a title of "first" or "last."

references to first names and their use in various contexts

No Comments

No Known Activations