INDEX

Explanations

sharp and intense physical descriptors

oai_token-act-pair · gpt-3.5-turbo

references to the word "sharp" 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 @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Alive

-0.83

pty

-0.80

FORE

-0.77

ORED

-0.74

perty

-0.74

uthor

-0.73

YL

-0.72

osponsors

-0.70

BLE

-0.69

rella

-0.68

POSITIVE LOGITS

ened

1.42

ening

1.30

ener

1.09

est

1.07

ness

0.99

 edges

0.90

edo

0.87

 elbows

0.86

eners

0.85

 edged

0.84

Activations Density 0.040%

sharp and intense physical descriptors

references to the word "sharp" in various contexts

No Comments

No Known Activations