INDEX

Explanations

words related to defects or issues in various contexts

oai_token-act-pair · gpt-3.5-turbo

terms associated with defects and deficiencies

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

zer

-0.76

verson

-0.68

sta

-0.67

fare

-0.67

corn

-0.65

cas

-0.64

ãĥĥãĤ¯

-0.64

brance

-0.63

zees

-0.63

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.63

POSITIVE LOGITS

ively

0.86

 defect

0.83

ions

0.82

smanship

0.78

 defects

0.77

icient

0.77

inia

0.76

ibility

0.76

itives

0.75

ors

0.74

Activations Density 0.013%

words related to defects or issues in various contexts

terms associated with defects and deficiencies

No Comments

No Known Activations