INDEX

Explanations

instances of the word "when" indicating temporal contexts

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.0.hook_mlp_out

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ographic

-1.50

begin

-1.44

 Firstly

-1.41

schild

-1.38

 parameter

-1.38

avier

-1.34

 Publishers

-1.34

 /**<

-1.33

cluding

-1.33

 certificate

-1.33

POSITIVE LOGITS

Ļª

4.49

±

4.21

ĨĴ

4.13

¢

4.10

¸

3.96

Ĳ

3.90

¾

3.75

µ

3.73

↵↵

3.73

Activations Density 0.354%

instances of the word "when" indicating temporal contexts

No Comments

No Known Activations