INDEX

Explanations

negative constructions related to inability or refusal

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

cannot handle

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

MessageTagHelper

-0.50

 Grüße

-0.46

 Grüsse

-0.45

GraphicsUnit

-0.42

 sensaciones

-0.41

 незавершена

-0.41

AnchorStyles

-0.40

Personendaten

-0.40

 pinulongan

-0.39

 AssemblyTitle

-0.39

POSITIVE LOGITS

 toler

0.59

 Handle

0.58

 HANDLE

0.57

 Handles

0.55

 handle

0.55

Handle

0.54

handle

0.54

 handles

0.53

 handling

0.52

 Toler

0.52

Activations Density 0.035%

negative constructions related to inability or refusal

cannot handle

No Comments

No Known Activations