INDEX

Explanations

contractions denoting possession (e.g., 's)

oai_token-act-pair · gpt-3.5-turbo

references to a specific year

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

folios

-0.81

itcher

-0.80

athy

-0.79

wright

-0.76

ships

-0.74

VO

-0.74

fo

-0.71

OTO

-0.70

eno

-0.69

FILE

-0.69

POSITIVE LOGITS

 edition

1.00

 inaugural

0.91

 incarnation

0.87

 installment

0.83

 iteration

0.82

 biggest

0.82

 anniversary

0.79

 winners

0.77

 editions

0.76

 festivities

0.75

Activations Density 0.055%

contractions denoting possession (e.g., 's)

references to a specific year

No Comments

No Known Activations