INDEX

Explanations

phrases related to criminal suspects or investigations

oai_token-act-pair · gpt-3.5-turbo

mentions of suspects in the context of criminal investigations

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

Not in Any Lists

No Comments

Negative Logits

ammy

-0.70

ann

-0.67

nces

-0.66

oted

-0.66

ŃĶ

-0.65

natal

-0.64

antic

-0.63

ntil

-0.63

touch

-0.62

tem

-0.62

POSITIVE LOGITS

 suspects

1.25

 suspect

1.00

 suspected

0.70

arnaev

0.70

CSI

0.69

pots

0.68

 offenders

0.67

istani

0.66

 culprit

0.65

 perpetrators

0.65

Activations Density 0.004%

phrases related to criminal suspects or investigations

mentions of suspects in the context of criminal investigations

No Comments

No Known Activations