INDEX

Explanations

references to the Vancouver Sedins, likely in a context related to ice hockey

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Sed" 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 @ 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

#$

-0.83

boards

-0.70

enegger

-0.65

cake

-0.64

foundation

-0.64

SHIP

-0.64

strap

-0.63

theless

-0.62

ãĥ£

-0.61

Hop

-0.61

POSITIVE LOGITS

uctive

1.20

entary

1.19

gew

1.13

imentary

1.09

iments

1.03

uction

0.98

ridor

0.97

ition

0.96

uct

0.91

Sed

0.89

Activations Density 0.023%

references to the Vancouver Sedins, likely in a context related to ice hockey

occurrences of the term "Sed" in various contexts

No Comments

No Known Activations