INDEX

Explanations

4-digit years

oai_token-act-pair · gpt-3.5-turbo

years or dates mentioned in the text

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

 speedy

-0.56

ggles

-0.55

 intangible

-0.55

verages

-0.54

 norm

-0.51

ube

-0.51

 continu

-0.50

Tags

-0.49

 Meter

-0.49

 substance

-0.48

POSITIVE LOGITS

-'

1.08

 alone

0.87

staking

0.85

\.

0.78

 when

0.74

0.70

å¹

0.69

when

0.69

âĢ²

0.69

 after

0.68

Activations Density 0.123%

4-digit years

years or dates mentioned in the text

No Comments

No Known Activations