INDEX

Explanations

words related to security breaches or unauthorized access

oai_token-act-pair · gpt-3.5-turbo

terms related to intrusions and security breaches

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ndra

-0.97

å§«

-0.90

creen

-0.89

wagen

-0.88

IELD

-0.83

baum

-0.82

hyde

-0.73

hare

-0.70

BALL

-0.70

yards

-0.68

POSITIVE LOGITS

uders

1.06

uding

1.05

 intr

1.01

inx

0.98

usions

0.93

igent

0.91

ention

0.89

iencies

0.87

uded

0.86

agg

0.85

Activations Density 0.008%

words related to security breaches or unauthorized access

terms related to intrusions and security breaches

No Comments

No Known Activations