INDEX

Explanations

references to written scripts or instructions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "script" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

hoff

-0.64

ktop

-0.64

 Gallup

-0.64

Dew

-0.63

 Jenner

-0.62

coat

-0.61

 Cheong

-0.60

 tide

-0.60

hm

-0.59

Gri

-0.58

POSITIVE LOGITS

ural

1.22

writers

1.07

writer

1.05

writing

0.99

urally

0.95

Pub

0.90

abal

0.83

 interpreter

0.82

ulator

0.79

smanship

0.79

Activations Density 0.029%

references to written scripts or instructions

occurrences of the word "script" in various contexts

No Comments

No Known Activations