INDEX

Explanations

phrases related to references or quotations in text

oai_token-act-pair · gpt-3.5-turbo

possessive forms and related possessive language

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

vg

-0.70

Ļ½

-0.68

endor

-0.64

rador

-0.62

amiya

-0.62

clipse

-0.62

ptives

-0.61

 Archangel

-0.61

ifier

-0.61

sted

-0.60

POSITIVE LOGITS

own

0.82

 penchant

0.73

 collective

0.72

 fingerprints

0.72

 sake

0.66

nai

0.66

ELF

0.66

cause

0.66

 birthplace

0.65

 guts

0.64

Activations Density 0.036%

phrases related to references or quotations in text

possessive forms and related possessive language

No Comments

No Known Activations