INDEX

Explanations

occurrences of the substring "th" in words

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

th term

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

parsedMessage

-0.79

principalColumn

-0.77

 سكانية

-0.73

SceneManagement

-0.73

-0.70

[@BOS@]

-0.69

<unused14>

-0.69

<unused16>

-0.69

<unused8>

-0.69

<pad>

-0.69

POSITIVE LOGITS

th

0.96

TH

0.81

uth

0.71

Th

0.66

Th

0.66

ath

0.66

ith

0.66

ATH

0.65

hth

0.64

UTH

0.63

Activations Density 0.097%

occurrences of the substring "th" in words

th term

No Comments

No Known Activations