INDEX

Explanations

phrasal verbs

oai_token-act-pair · gpt-3.5-turbo

negations and topics related to rules or authority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Lov

-0.91

 Leban

-0.86

 Manz

-0.83

hl

-0.83

Kiw

-0.83

 Hass

-0.82

Kir

-0.81

Kid

-0.81

 Beir

-0.79

coh

-0.79

POSITIVE LOGITS

atum

1.00

ativity

0.99

ato

0.99

ance

0.93

ACY

0.92

ata

0.91

atus

0.87

ure

0.87

utation

0.85

URE

0.85

Activations Density 0.232%

phrasal verbs

negations and topics related to rules or authority

No Comments

No Known Activations