INDEX

Explanations

words related to high resolution or detailed images

oai_token-act-pair · gpt-3.5-turbo

terms related to image resolution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

redo

-0.98

bara

-0.90

erd

-0.75

opl

-0.73

arus

-0.72

tail

-0.71

obin

-0.71

spring

-0.70

tails

-0.70

uana

-0.68

POSITIVE LOGITS

 resolution

1.11

 resolutions

1.02

 Resolution

1.00

olution

0.84

 mismatch

0.82

olutions

0.79

0.77

 sickness

0.76

 sensitivity

0.75

Activations Density 0.015%

words related to high resolution or detailed images

terms related to image resolution

No Comments

No Known Activations