INDEX

Explanations

words related to physical features or characteristics, specifically focusing on wings

oai_token-act-pair · gpt-3.5-turbo

references to wings, indicating focus on flight or winged creatures and concepts

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

ALLY

-0.80

 Cumber

-0.76

ICAN

-0.71

 Barnett

-0.71

OWN

-0.68

 Scotia

-0.68

UTE

-0.65

çĶ°

-0.64

 Identity

-0.64

 Holmes

-0.64

POSITIVE LOGITS

pan

1.30

uit

1.23

pread

1.20

uits

1.13

wings

1.09

 wings

1.09

hops

1.03

hip

1.02

pots

1.00

layer

0.93

Activations Density 0.025%

words related to physical features or characteristics, specifically focusing on wings

references to wings, indicating focus on flight or winged creatures and concepts

No Comments

No Known Activations