INDEX

Explanations

paths and URL patterns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Cor

-0.65

NamedQueries

-0.60

ˌ

-0.57

 приходи

-0.57

 niks

-0.57

 autorisation

-0.57

Preparazione

-0.56

eiro

-0.56

巴士

-0.56

gib

-0.55

POSITIVE LOGITS

(__('

0.82

 poffible

0.78

('--

0.75

('../

0.75

thodon

0.74

 Majefty

0.74

 ſche

0.73

 pleaſure

0.73

 occaf

0.73

Activations Density 0.089%

paths and URL patterns

No Comments

No Known Activations