INDEX

Explanations

locations or settings within a narrative

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "place" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

xf

-0.65

ovich

-0.65

hart

-0.63

hawks

-0.62

ICO

-0.62

ili

-0.61

scl

-0.59

 Tayyip

-0.59

Serial

-0.59

 Aviv

-0.58

POSITIVE LOGITS

 place

3.77

place

2.52

 Place

2.35

Place

2.19

 places

2.12

places

1.82

 Places

1.62

 spot

1.48

 location

1.40

 position

1.28

Activations Density 0.038%

locations or settings within a narrative

occurrences of the word "place" in various contexts

No Comments

No Known Activations