INDEX

Explanations

phrases indicating simultaneous occurrence or actions happening at the same time

oai_token-act-pair · gpt-3.5-turbo

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

ntil

-0.74

uable

-0.69

pmwiki

-0.67

avorite

-0.67

efully

-0.67

prus

-0.66

perm

-0.66

ilater

-0.65

ggles

-0.65

nce

-0.65

POSITIVE LOGITS

0.70

as

0.70

 respecting

0.69

we

0.63

.............

0.61

shapeshifter

0.60

 that

0.58

 they

0.57

 acknowledging

0.57

 emphasizing

0.57

Activations Density 0.031%

phrases indicating simultaneous occurrence or actions happening at the same time

No Comments

No Known Activations