INDEX

Explanations

fragments of dialogue or testimony in legal contexts

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

mistake reflection

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

AnchorStyles

-0.69

AddTagHelper

-0.56

rungsseite

-0.54

MessageState

-0.52

DockStyle

-0.50

 snippetHide

-0.49

parsedMessage

-0.49

Personensuche

-0.49

########.

-0.49

CppMethod

-0.48

POSITIVE LOGITS

 hindsight

0.44

 fault

0.40

 negligence

0.40

 blame

0.38

тельность

0.38

 mistake

0.38

 carelessness

0.37

反省

0.35

 blaming

0.35

 stupid

0.35

Activations Density 0.081%

fragments of dialogue or testimony in legal contexts

mistake reflection

No Comments

No Known Activations