INDEX

Explanations

pronouns (primarily "His") followed by specific descriptions or actions

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual's experiences or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

dding

-0.70

å¥³

-0.69

fitting

-0.66

ÙĲ

-0.66

qi

-0.62

ãĥ´ãĤ¡

-0.61

Õ

-0.61

Ð¼

-0.60

Ðµ

-0.60

Ð¸

-0.59

POSITIVE LOGITS

panic

1.01

 Majesty

0.95

own

0.94

Own

0.94

resy

0.93

self

0.89

itage

0.84

anmar

0.84

 millenn

0.82

rera

0.81

Activations Density 0.025%

pronouns (primarily "His") followed by specific descriptions or actions

references to a specific individual's experiences or actions

No Comments

No Known Activations