INDEX

Explanations

locations or specific places, particularly mountains such as Mt. Everest

oai_token-act-pair · gpt-3.5-turbo

references to specific mountains and locations associated with them

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

 Dragonbound

-0.85

Interstitial

-0.80

 IMAGES

-0.75

SHIP

-0.73

 briefs

-0.67

thro

-0.66

BACK

-0.65

interstitial

-0.64

makers

-0.64

served

-0.63

POSITIVE LOGITS

awar

1.17

emonic

1.01

unia

0.89

oday

0.82

olt

0.81

ains

0.81

umble

0.81

acular

0.81

iami

0.80

gt

0.80

Activations Density 0.012%

locations or specific places, particularly mountains such as Mt. Everest

references to specific mountains and locations associated with them

No Comments

No Known Activations