INDEX

Explanations

attends to negation-related tokens from their affirmative counterparts

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.59

2:0.11

3:0.02

4:0.06

5:0.05

6:0.05

7:0.06

Negative Logits

 &___

-0.96

sizeCache

-0.90

RegressionTest

-0.85

 disambiguazione

-0.85

tagHelperRunner

-0.85

principalColumn

-0.83

LookAnd

-0.82

 propOrder

-0.82

 CreateTagHelper

-0.81

 CURIAM

-0.80

POSITIVE LOGITS

0.54

0.53

0.48

0.47

0.46

0.45

The

0.45

0.43

In

0.42

Activations Density 0.096%

attends to negation-related tokens from their affirmative counterparts

No Comments

No Known Activations