INDEX

Explanations

words related to negative conditions or situations, such as overcrowding, danger, and illness

oai_token-act-pair · gpt-3.5-turbo

words related to overcrowding and mental distress

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

 Haram

-0.72

uana

-0.65

otine

-0.60

Downloadha

-0.59

 breath

-0.58

FORE

-0.58

 clinch

-0.58

 Imam

-0.56

 Webster

-0.56

 breathe

-0.54

POSITIVE LOGITS

ed

2.06

edly

1.58

ing

1.37

edIn

1.35

eds

1.10

ships

1.09

eded

1.07

arily

1.05

eding

1.02

esy

1.02

Activations Density 0.045%

words related to negative conditions or situations, such as overcrowding, danger, and illness

words related to overcrowding and mental distress

No Comments

No Known Activations