INDEX

Explanations

mentions of the word "giant."

oai_token-act-pair · gpt-3.5-turbo

repetitive occurrences of the term "giant."

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

IFrame

Link

Negative Logits

earchers

-0.90

skirts

-0.75

uberty

-0.72

cffffcc

-0.70

atever

-0.69

agan

-0.68

omsky

-0.67

Downloadha

-0.66

upuncture

-0.64

MER

-0.63

POSITIVE LOGITS

iant

1.07

ieth

0.99

ially

0.91

yne

0.84

ãĥ¼ãĥĨãĤ£

0.84

riad

0.82

iful

0.75

ial

0.73

iants

0.71

ifully

0.70

Activations Density 0.007%

mentions of the word "giant."

repetitive occurrences of the term "giant."

No Known Activations