INDEX

Explanations

locations or geographical entities

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and titles related to organizations, locations, and people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 ..."

-0.79

akable

-0.75

Ì¶

-0.65

...)

-0.64

 imaginable

-0.62

 thereof

-0.60

Ò

-0.59

NVIDIA

-0.58

Zip

-0.57

POSITIVE LOGITS

 spokesman

1.09

 spokeswoman

1.08

 meanwhile

1.02

 spokesperson

0.97

 airs

0.92

 operates

0.88

 boasts

0.87

 responded

0.86

 also

0.86

 declined

0.85

Activations Density 0.480%

locations or geographical entities

proper nouns, particularly names and titles related to organizations, locations, and people

No Comments

No Known Activations