INDEX

Explanations

phrases asking about actions or states

oai_token-act-pair · gpt-3.5-turbo

questions that begin with "What are."

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

oire

-0.81

dom

-0.81

speak

-0.78

icism

-0.78

osate

-0.76

town

-0.74

iasis

-0.71

ittered

-0.69

matter

-0.69

 Verse

-0.68

POSITIVE LOGITS

senal

1.06

wolves

1.00

wolf

0.85

nt

0.82

nda

0.81

 gonna

0.72

nd

0.71

 legion

0.68

 hereby

0.67

 Ãľ

0.65

Activations Density 0.128%

phrases asking about actions or states

questions that begin with "What are."

No Comments

No Known Activations