INDEX

Explanations

phrases containing the word "are" that are unrelated and lack a clear cohesive theme

oai_token-act-pair · gpt-3.5-turbo

phrases indicating existence or presence, particularly in relation to entities or actions

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

krit

-0.68

ienne

-0.68

mberg

-0.65

uria

-0.65

enhagen

-0.65

tal

-0.64

javascript

-0.64

chance

-0.62

itiz

-0.61

iT

-0.61

POSITIVE LOGITS

 respectively

2.02

 alike

1.75

depending

1.01

 together

1.01

 jointly

0.97

 collide

0.93

together

0.90

 respective

0.88

 interchangeable

0.84

 simultaneously

0.83

Activations Density 0.850%

phrases containing the word "are" that are unrelated and lack a clear cohesive theme

phrases indicating existence or presence, particularly in relation to entities or actions

No Comments

No Known Activations