INDEX

Explanations

verbs or phrases related to making decisions or recommendations

oai_token-act-pair · gpt-3.5-turbo

statements that indicate stated opinions or beliefs

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

isal

-0.75

cation

-0.70

mania

-0.62

asus

-0.60

panic

-0.60

ommel

-0.59

aml

-0.59

lance

-0.59

tery

-0.59

etheless

-0.58

POSITIVE LOGITS

 themselves

1.19

 selves

0.97

selves

0.84

 helmets

0.68

 jointly

0.68

 individually

0.66

 uniforms

0.65

 mouths

0.64

 microphones

0.64

aughtered

0.64

Activations Density 0.980%

verbs or phrases related to making decisions or recommendations

statements that indicate stated opinions or beliefs

No Comments

No Known Activations