INDEX

Explanations

references to a specific named person or entity (Higgs) in varied contexts

oai_token-act-pair · gpt-3.5-turbo

references to the Higgs boson or related particles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 captured

-0.73

Bed

-0.69

 authorized

-0.66

 provincial

-0.65

cam

-0.64

 resort

-0.64

Ara

-0.64

 Palm

-0.63

 Chart

-0.62

 vacation

-0.62

POSITIVE LOGITS

iggs

4.50

igg

1.51

iggins

1.32

utherford

1.08

ickey

1.02

allas

1.00

essler

0.97

atton

0.96

ourke

0.95

iggle

0.94

Activations Density 0.017%

references to a specific named person or entity (Higgs) in varied contexts

references to the Higgs boson or related particles

No Comments

No Known Activations