INDEX

Explanations

terms related to loss or negative impact

oai_token-act-pair · gpt-3.5-turbo

phrases centered around the concept of loss

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ucci

-0.82

çīĪ

-0.82

nis

-0.76

ulhu

-0.75

 Cosponsors

-0.75

iquette

-0.73

eller

-0.71

KER

-0.70

soType

-0.70

leground

-0.70

POSITIVE LOGITS

 virginity

0.82

 unwanted

0.74

 consciousness

0.71

 tack

0.71

 Crimea

0.71

 custody

0.70

 interest

0.70

 limbs

0.70

 control

0.69

luster

0.68

Activations Density 0.099%

terms related to loss or negative impact

phrases centered around the concept of loss

No Comments

No Known Activations