INDEX

Explanations

phrases related to political figures, particularly Donald Trump

oai_token-act-pair · gpt-3.5-turbo

instances of a specific character or symbol being used in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 guiActiveUnfocused

-0.86

 scattering

-0.79

 scatter

-0.77

 decomp

-0.76

 confinement

-0.74

 unmarked

-0.74

 Nept

-0.70

 microscope

-0.70

 Guinness

-0.68

 limitation

-0.67

POSITIVE LOGITS

¹

1.03

į

0.97

Republicans

0.92

Joe

0.91

Obama

0.91

Rick

0.91

Politics

0.91

Į

0.90

CNN

0.89

²

0.88

Activations Density 0.563%

phrases related to political figures, particularly Donald Trump

instances of a specific character or symbol being used in the text

No Comments

No Known Activations