INDEX

Explanations

expressions indicating reaching a high or extreme level or taking drastic actions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating limits or boundaries in actions or ideas

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

itu

-0.76

bed

-0.68

ivered

-0.65

 Sovere

-0.65

ipation

-0.63

ixture

-0.63

Nep

-0.62

ukes

-0.61

stable

-0.59

ixtures

-0.59

POSITIVE LOGITS

WARD

0.89

 sidx

0.82

 overboard

0.81

 derog

0.79

 towards

0.75

 unnoticed

0.74

irtual

0.74

 toward

0.74

 boldly

0.72

 lengths

0.68

Activations Density 0.057%

expressions indicating reaching a high or extreme level or taking drastic actions

phrases indicating limits or boundaries in actions or ideas

No Comments

No Known Activations