INDEX

Explanations

proper nouns related to people or places

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "bot."

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

pees

-0.78

eah

-0.72

wd

-0.64

 Enix

-0.62

arp

-0.62

udos

-0.61

pee

-0.60

vel

-0.60

leased

-0.60

 creeps

-0.59

POSITIVE LOGITS

 Reloaded

0.81

ãĤ´ãĥ³

0.80

enegger

0.79

ertodd

0.74

inez

0.71

merce

0.70

Connector

0.69

REDACTED

0.69

illation

0.65

opia

0.63

Activations Density 0.046%

proper nouns related to people or places

occurrences of the term "bot."

No Comments

No Known Activations