INDEX

Explanations

words related to high places or elevation

oai_token-act-pair · gpt-3.5-turbo

locations or references to specific places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enz

-0.74

anish

-0.68

 inval

-0.64

icter

-0.63

requ

-0.62

oice

-0.62

 proportion

-0.61

ubi

-0.60

 fiance

-0.60

 genetically

-0.59

POSITIVE LOGITS

tyard

1.06

 Corridor

1.05

 corridor

1.05

 corridors

0.99

 overlooking

0.97

 courtyard

0.93

 cavern

0.90

 chapel

0.87

 Hotel

0.87

 hallway

0.86

Activations Density 1.142%

words related to high places or elevation

locations or references to specific places

No Comments

No Known Activations