INDEX

Explanations

words ending in 'ed'

oai_token-act-pair · gpt-3.5-turbo

the past tense form of verbs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

å¥³

-0.86

FTWARE

-0.79

 Leilan

-0.78

å£

-0.78

ORY

-0.71

ãĥ¯ãĥ³

-0.71

åħī

-0.68

itars

-0.67

BOOK

-0.64

 caliber

-0.64

POSITIVE LOGITS

ict

0.89

nesday

0.89

uct

0.88

ging

0.83

uled

0.83

own

0.83

dit

0.81

ded

0.80

usa

0.79

ding

0.79

Activations Density 0.072%

words ending in 'ed'

the past tense form of verbs

No Comments

No Known Activations