INDEX

Explanations

people's jobs or roles within different scenarios

oai_token-act-pair · gpt-3.5-turbo

terms related to individuals in various roles or occupations

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

 Dates

-0.72

ernels

-0.71

 Pieces

-0.70

 Enemies

-0.68

 calendars

-0.67

ories

-0.66

headers

-0.65

endars

-0.65

ippers

-0.65

 Trails

-0.65

POSITIVE LOGITS

 named

1.14

 testified

1.04

 wrote

0.97

 remarked

0.93

 surn

0.92

who

0.91

 told

0.90

 nicknamed

0.89

 likened

0.85

 threw

0.85

Activations Density 0.244%

people's jobs or roles within different scenarios

terms related to individuals in various roles or occupations

No Comments

No Known Activations