INDEX

Explanations

adjectives or past participles ending in '-ed' or '-ing'

oai_token-act-pair · gpt-3.5-turbo

references to pie or pie-related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iveness

-0.79

ians

-0.75

runners

-0.70

 Examiner

-0.70

worthiness

-0.69

sers

-0.69

ulk

-0.68

afety

-0.67

interrupted

-0.66

nesses

-0.63

POSITIVE LOGITS

ced

1.19

cing

1.16

 pies

1.12

 crust

1.12

pie

1.11

pie

1.08

Pie

0.96

ces

0.95

Pie

0.93

wich

0.93

Activations Density 0.032%

adjectives or past participles ending in '-ed' or '-ing'

references to pie or pie-related concepts

No Comments

No Known Activations