INDEX

Explanations

phrases or terms related to press releases or news articles with a focus on legal or political issues

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

HuggingFaceFW/fineweb

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 grow

-0.75

 deal

-0.75

if

-0.73

 tend

-0.73

 keep

-0.72

 enjoy

-0.72

am

-0.72

 bring

-0.71

 even

-0.71

 pick

-0.71

POSITIVE LOGITS

 affor

2.25

 maneu

2.11

fta

2.10

ftu

2.10

 emphat

2.08

 stockholm

2.08

fup

2.02

 excru

2.02

 increa

2.02

 wien

2.00

Activations Density 0.210%

phrases or terms related to press releases or news articles with a focus on legal or political issues

No Comments

No Known Activations