INDEX

Explanations

words related to actions that have been completed or terminated

oai_token-act-pair · gpt-3.5-turbo

words that convey a sense of a process or state of being

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

FTWARE

-0.70

dfx

-0.68

 glim

-0.65

ood

-0.64

Ey

-0.61

 Beans

-0.60

Availability

-0.59

 Osiris

-0.56

earchers

-0.56

 rabbi

-0.56

POSITIVE LOGITS

ating

0.87

aukee

0.83

htaking

0.80

SHIP

0.78

anamo

0.76

onite

0.76

rome

0.76

uated

0.76

urally

0.76

atable

0.75

Activations Density 0.046%

words related to actions that have been completed or terminated

words that convey a sense of a process or state of being

No Comments

No Known Activations