INDEX

Explanations

references to classified information

oai_token-act-pair · gpt-3.5-turbo

references to classified information and documents

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

annis

-0.77

oller

-0.77

LV

-0.71

à

-0.70

Flor

-0.68

Vij

-0.68

issance

-0.67

begin

-0.65

Il

-0.65

alter

-0.64

POSITIVE LOGITS

 briefings

1.11

 papers

0.89

 documents

0.88

classified

0.86

 cables

0.84

oslov

0.83

 markings

0.83

 classified

0.82

abad

0.81

 material

0.81

Activations Density 0.017%

references to classified information

references to classified information and documents

No Comments

No Known Activations