INDEX

Explanations

mentions of significant achievements, such as awards, publications, and accomplishments

oai_token-act-pair · gpt-3.5-turbo

nouns related to actions, achievements, and roles

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

ensis

-0.67

 Examiner

-0.66

 Fury

-0.64

 Tactics

-0.61

 Redemption

-0.60

arde

-0.60

-0.59

 Survival

-0.59

 Doodle

-0.58

POSITIVE LOGITS

 including

1.13

 simultaneously

1.03

 besides

1.00

 ranging

0.99

 imaginable

0.97

hips

0.93

 throughout

0.91

 relating

0.89

 totaling

0.87

 apiece

0.87

Activations Density 0.461%

mentions of significant achievements, such as awards, publications, and accomplishments

nouns related to actions, achievements, and roles

No Comments

No Known Activations