INDEX

Explanations

phrases related to extremist groups or terrorists

oai_token-act-pair · gpt-3.5-turbo

references to jihadist individuals or groups

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

ances

-0.72

 Holt

-0.70

--+

-0.70

 Plant

-0.69

NRS

-0.67

ospital

-0.67

rost

-0.67

ANC

-0.65

clip

-0.65

oro

-0.65

POSITIVE LOGITS

 jihadists

1.04

 jihadist

1.02

 guiActiveUn

0.95

 terrorists

0.86

 jihad

0.83

emort

0.82

 terrorism

0.81

querque

0.81

apons

0.81

 terror

0.80

Activations Density 0.005%

phrases related to extremist groups or terrorists

references to jihadist individuals or groups

No Comments

No Known Activations