INDEX

Explanations

political and legal terms

oai_token-act-pair · gpt-3.5-turbo

mentions of legal and governmental proceedings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 eleph

-0.88

 ende

-0.85

 citiz

-0.85

 crumble

-0.80

 undecided

-0.79

icter

-0.79

 tram

-0.78

 waterfall

-0.78

 landsl

-0.78

 offline

-0.77

POSITIVE LOGITS

Interested

1.92

According

1.77

The

1.69

Speaking

1.67

Beginning

1.61

Following

1.61

During

1.60

In

1.59

Officials

1.59

Instead

1.57

Activations Density 0.269%

political and legal terms

mentions of legal and governmental proceedings

No Comments

No Known Activations