INDEX

Explanations

names of organizations and companies

oai_token-act-pair · gpt-3.5-turbo

corporate affiliations and organizational roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 chase

-0.74

 rush

-0.73

 trope

-0.70

 silence

-0.70

 routine

-0.69

 headlines

-0.68

jriwal

-0.68

 excuse

-0.67

 coughing

-0.66

 background

-0.66

POSITIVE LOGITS

 Associates

0.87

 Ventures

0.86

Works

0.85

 Strategies

0.85

 Productions

0.82

Bio

0.80

eworks

0.76

 Artists

0.75

Choice

0.75

Mind

0.75

Activations Density 0.441%

names of organizations and companies

corporate affiliations and organizational roles

No Comments

No Known Activations