INDEX

Explanations

phrases instructing or reminding about staying informed or updated

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize ongoing action or reminders

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

é¾įåĸļå£«

-0.77

ãĥ¼ãĥĨ

-0.72

ELF

-0.70

ãĥ£

-0.67

éĹ

-0.66

ENTS

-0.65

cffffcc

-0.63

tom

-0.63

 magically

-0.63

Jub

-0.62

POSITIVE LOGITS

watch

0.87

 Keep

0.87

Keep

0.81

 Your

0.79

 Recall

0.78

ership

0.76

ers

0.75

chin

0.74

Skip

0.74

gon

0.74

Activations Density 0.027%

phrases instructing or reminding about staying informed or updated

phrases that emphasize ongoing action or reminders

No Comments

No Known Activations