INDEX

Explanations

instances of the word "Victoria" with varying intensities

oai_token-act-pair · gpt-3.5-turbo

references to the name "Victoria."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

earance

-0.81

hest

-0.81

etary

-0.81

bler

-0.78

edly

-0.78

mble

-0.75

nesses

-0.74

backer

-0.74

ystem

-0.73

viation

-0.72

POSITIVE LOGITS

 Beckham

1.03

 Falls

0.82

 Fleming

0.75

 Slater

0.73

uates

0.72

 Police

0.71

isle

0.71

icum

0.71

 Broadcasting

0.70

 Gardens

0.70

Activations Density 0.018%

instances of the word "Victoria" with varying intensities

references to the name "Victoria."

No Comments

No Known Activations