INDEX

Explanations

names and entities in a variety of contexts

oai_token-act-pair · gpt-3.5-turbo

unique identifiers or names

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Ken

-0.91

eller

-0.87

 Kenyan

-0.87

Miller

-0.86

apo

-0.86

 Miller

-0.84

ale

-0.84

 Aliens

-0.83

 Kenya

-0.83

Ken

-0.83

POSITIVE LOGITS

Str

1.33

STR

1.24

 string

1.21

 String

1.21

string

1.19

 Bridge

1.19

String

1.17

 strings

1.14

str

1.14

STR

1.14

Activations Density 0.374%

names and entities in a variety of contexts

unique identifiers or names

No Comments

No Known Activations