INDEX

Explanations

words related to danger and medical conditions ending in '-ous'

oai_token-act-pair · gpt-3.5-turbo

words with the suffix "-rous" often related to describing characteristics or qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

NPR

-0.75

ŃĶ

-0.71

 WARN

-0.71

ahi

-0.71

 Loan

-0.69

çĶ°

-0.69

opsy

-0.67

HOME

-0.64

 stump

-0.64

GET

-0.63

POSITIVE LOGITS

rous

1.22

atile

1.12

theless

1.06

rontal

0.98

atility

0.90

 oxide

0.86

ious

0.83

entimes

0.82

Magikarp

0.78

phrine

0.74

Activations Density 0.010%

words related to danger and medical conditions ending in '-ous'

words with the suffix "-rous" often related to describing characteristics or qualities

No Comments

No Known Activations