INDEX

Explanations

the word "bag" with a high activation level

oai_token-act-pair · gpt-3.5-turbo

references to bags

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

vironment

-0.72

hower

-0.67

terday

-0.66

 Galile

-0.65

issance

-0.65

 relations

-0.64

 spect

-0.62

 preschool

-0.61

nesota

-0.60

 spectrum

-0.60

POSITIVE LOGITS

ging

1.23

bag

1.19

gie

1.15

gery

1.13

ged

1.12

bags

1.10

pipe

1.05

glers

0.99

rill

0.96

gers

0.96

Activations Density 0.016%

the word "bag" with a high activation level

references to bags

No Comments

No Known Activations