INDEX

Explanations

phrases related to hacking or cyber attacks

oai_token-act-pair · gpt-3.5-turbo

words related to hacking or unauthorized access

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Debor

-0.92

psons

-0.69

ocol

-0.69

stros

-0.68

olith

-0.68

 Krug

-0.67

 Niet

-0.67

teness

-0.65

ieves

-0.64

iations

-0.63

POSITIVE LOGITS

acked

1.13

ACK

1.00

acks

0.99

acking

0.89

intosh

0.85

lash

0.83

luster

0.77

acker

0.75

enzie

0.73

innon

0.72

Activations Density 0.012%

phrases related to hacking or cyber attacks

words related to hacking or unauthorized access

No Comments

No Known Activations