INDEX

Explanations

instances of the word "probe."

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "probe" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 entitled

-0.71

 Calvin

-0.67

 Social

-0.66

 Alley

-0.64

 Towns

-0.63

styles

-0.63

 ILCS

-0.63

 Languages

-0.63

Social

-0.61

yards

-0.61

POSITIVE LOGITS

 probe

3.85

 probes

3.07

 Probe

2.62

 probing

2.20

 investigation

1.75

 inquiry

1.62

 investigate

1.38

 investigations

1.36

 regulator

1.29

 investigator

1.29

Activations Density 0.013%

instances of the word "probe."

repeated mentions of the word "probe" in various contexts

No Comments

No Known Activations