INDEX

Explanations

phrases including the word "Tale."

oai_token-act-pair · gpt-3.5-turbo

references to specific narratives or stories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

wm

-0.98

OWS

-0.76

essa

-0.69

Pl

-0.69

ipers

-0.68

 Flores

-0.67

=#

-0.66

 Shields

-0.66

pause

-0.66

ey

-0.65

POSITIVE LOGITS

 Tale

2.56

Associated

1.96

Jal

1.55

Hak

1.49

 Jama

1.42

 Tales

1.19

 Travels

0.94

 CCTV

0.90

Bam

0.87

 Barber

0.86

Activations Density 0.049%

phrases including the word "Tale."

references to specific narratives or stories

No Comments

No Known Activations