INDEX

Explanations

non-standard token patterns, possibly for a specific task or encoding scheme

oai_token-act-pair · gpt-3.5-turbo

instances of numerical data or quantitative references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ividual

-0.75

utsche

-0.73

hene

-0.72

 assurance

-0.70

 lifes

-0.69

abwe

-0.68

 subpoena

-0.68

earchers

-0.68

 liberty

-0.68

ertodd

-0.67

POSITIVE LOGITS

Trivia

1.62

Characters

1.60

Contents

1.45

Appearances

1.44

pmwiki

1.37

Gallery

1.28

Gameplay

1.23

Examples

1.19

References

1.19

âĶĢâĶĢâĶĢâĶĢ

1.18

Activations Density 0.329%

non-standard token patterns, possibly for a specific task or encoding scheme

instances of numerical data or quantitative references

No Comments

No Known Activations