INDEX

Explanations

phrases indicating a negative consequence or impact

oai_token-act-pair · gpt-3.5-turbo

phrases indicating abandonment or loss

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

wcsstore

-0.78

buster

-0.67

 Refer

-0.62

wana

-0.62

intend

-0.61

Refer

-0.60

percent

-0.60

 relayed

-0.59

breaker

-0.59

inance

-0.59

POSITIVE LOGITS

 undone

1.03

 scars

0.95

overs

0.93

 unanswered

0.87

 behind

0.85

 unexpl

0.82

 footprints

0.79

 untold

0.79

 unexplained

0.79

him

0.77

Activations Density 0.035%

phrases indicating a negative consequence or impact

phrases indicating abandonment or loss

No Comments

No Known Activations