INDEX

Explanations

phrases emphasizing correctness or completion

oai_token-act-pair · gpt-3.5-turbo

affirmations or confirmations within statements

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

 Krug

-0.78

moil

-0.76

ngth

-0.74

 Agric

-0.63

alks

-0.62

ras

-0.62

hester

-0.60

cius

-0.60

clock

-0.59

plates

-0.59

POSITIVE LOGITS

SPONSORED

0.81

ECA

0.69

 coincidence

0.68

natureconservancy

0.68

UGH

0.65

POS

0.65

PLIED

0.64

ï¸

0.63

 scary

0.63

soType

0.62

Activations Density 0.298%

phrases emphasizing correctness or completion

affirmations or confirmations within statements

No Comments

No Known Activations