INDEX

Explanations

phrases related to specific names or entities, such as "Philip P. DiSte"

oai_token-act-pair · gpt-3.5-turbo

suffixes and endings of words, particularly focusing on affixes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 soph

-0.63

ob

-0.62

abama

-0.61

achelor

-0.61

 caliber

-0.60

 Spears

-0.60

idi

-0.60

 inco

-0.59

 Tanz

-0.59

paralle

-0.59

POSITIVE LOGITS

ings

1.21

cation

1.03

ment

1.00

ing

0.99

ingly

0.98

ees

0.96

ables

0.93

ments

0.93

ance

0.91

atives

0.88

Activations Density 0.209%

phrases related to specific names or entities, such as "Philip P. DiSte"

suffixes and endings of words, particularly focusing on affixes

No Comments

No Known Activations