INDEX

Explanations

instances of the word 'said' followed by a descriptive word

oai_token-act-pair · gpt-3.5-turbo

instances of the word "said" and its variants

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

ntil

-0.74

xtap

-0.70

potion

-0.70

 ILCS

-0.67

Ranked

-0.67

EDIT

-0.67

poon

-0.66

plete

-0.66

oufl

-0.65

pleting

-0.65

POSITIVE LOGITS

 goodbye

0.92

doms

0.74

 earlier

0.73

 sarcast

0.72

 anecd

0.69

 afterward

0.67

sonian

0.65

 aloud

0.65

 respondent

0.63

 afterwards

0.62

Activations Density 0.138%

instances of the word 'said' followed by a descriptive word

instances of the word "said" and its variants

No Comments

No Known Activations