INDEX

Explanations

mentions of the word "ant"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "ant" 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

UGE

-0.67

 Tens

-0.64

 Erie

-0.60

forth

-0.59

OME

-0.58

yy

-0.58

 Flames

-0.58

 Shutdown

-0.56

 Newfoundland

-0.56

ï¸ı

-0.55

POSITIVE LOGITS

emort

1.30

iques

1.29

elope

1.21

imony

1.13

iven

1.12

acid

1.11

icy

1.11

icol

1.07

onym

1.06

onyms

1.06

Activations Density 0.022%

mentions of the word "ant"

occurrences of the prefix "ant" in various contexts

No Comments

No Known Activations