INDEX

Explanations

phrases related to specific locations and properties, particularly ones with financial implications

oai_token-act-pair · gpt-3.5-turbo

negative sentiments associated with entities or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 abrupt

-0.72

EW

-0.67

 indo

-0.66

 accents

-0.66

 abruptly

-0.65

 cues

-0.65

 underscore

-0.64

 altogether

-0.64

 augmented

-0.63

 substitutes

-0.63

POSITIVE LOGITS

dozen

1.23

Lago

1.06

minute

1.05

lot

1.03

dos

1.03

Cola

1.01

service

1.01

cycle

1.00

week

1.00

rama

0.99

Activations Density 0.022%

phrases related to specific locations and properties, particularly ones with financial implications

negative sentiments associated with entities or situations

No Comments

No Known Activations