INDEX

Explanations

instances where a person or group is associated with another person or group

oai_token-act-pair · gpt-3.5-turbo

repetitions of the word "also"

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

ishers

-0.70

maxwell

-0.66

emies

-0.65

agus

-0.64

liest

-0.62

anwhile

-0.61

OPLE

-0.61

rarily

-0.60

orest

-0.59

orse

-0.59

POSITIVE LOGITS

 happens

0.86

 applies

0.85

 qualifies

0.83

 serves

0.79

 owns

0.77

 served

0.76

 referred

0.76

 benefited

0.75

 participated

0.75

 exists

0.74

Activations Density 0.102%

instances where a person or group is associated with another person or group

repetitions of the word "also"

No Comments

No Known Activations