INDEX

Explanations

the letter 't' occurring in various contexts

oai_token-act-pair · gpt-3.5-turbo

negative contractions related to the term "doesn't."

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

 Reloaded

-0.79

 Penguin

-0.77

 behavi

-0.75

Reviewer

-0.70

 Seas

-0.68

soType

-0.68

itiz

-0.68

è»

-0.68

 Passage

-0.66

 Communities

-0.64

POSITIVE LOGITS

 necessarily

1.11

 bother

0.99

cha

0.96

 anymore

0.88

ween

0.87

otally

0.85

weet

0.85

rave

0.83

aylor

0.83

 exactly

0.83

Activations Density 0.107%

the letter 't' occurring in various contexts

negative contractions related to the term "doesn't."

No Comments

No Known Activations