INDEX

Explanations

comparative terms indicating a superior or preferable option

oai_token-act-pair · gpt-3.5-turbo

comparative adjectives indicating superiority or quality

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

————

-0.76

Cola

-0.74

aren

-0.73

eva

-0.71

amera

-0.70

nikov

-0.69

ledge

-0.68

aven

-0.65

meg

-0.64

emi

-0.64

POSITIVE LOGITS

 portion

0.92

iest

0.90

 ones

0.90

 part

0.87

 sections

0.85

 side

0.84

 portions

0.81

 aspects

0.80

liest

0.80

 minded

0.76

Activations Density 0.145%

comparative terms indicating a superior or preferable option

comparative adjectives indicating superiority or quality

No Comments

No Known Activations