INDEX

Explanations

words related to supports or structures that are used to keep something in place or to provide assistance

oai_token-act-pair · gpt-3.5-turbo

the letter 'l' 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

SPONSORED

-0.83

aez

-0.77

Pax

-0.67

 Proced

-0.66

abyte

-0.65

cheon

-0.65

soon

-0.65

 Seym

-0.64

epad

-0.63

henko

-0.63

POSITIVE LOGITS

3.08

1.60

ll

1.41

1.36

1.29

1.28

lug

1.25

lo

1.25

li

1.21

le

1.20

Activations Density 0.016%

words related to supports or structures that are used to keep something in place or to provide assistance

the letter 'l' in various contexts

No Comments

No Known Activations