INDEX

Explanations

qualifiers indicating a high degree of emotion or physical state

oai_token-act-pair · gpt-3.5-turbo

intensifiers or modifiers expressing strong emotions or feelings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ð

-0.76

any

-0.71

 Combine

-0.69

çķ

-0.69

ulence

-0.69

ãĥ¯

-0.65

 Landing

-0.65

oras

-0.65

 Fail

-0.63

yles

-0.63

POSITIVE LOGITS

 pleased

1.52

 thankful

1.51

 grateful

1.49

 excited

1.44

 confident

1.44

 proud

1.42

 aware

1.40

 impressed

1.39

 thrilled

1.37

 nervous

1.36

Activations Density 0.164%

qualifiers indicating a high degree of emotion or physical state

intensifiers or modifiers expressing strong emotions or feelings

No Comments

No Known Activations