INDEX

Explanations

phrases indicating a purpose or intention

oai_token-act-pair · gpt-3.5-turbo

the conjunction "that" used to introduce clauses or explanations

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

ãĤ»

-0.65

ARK

-0.64

ocrates

-0.63

ahime

-0.63

atis

-0.62

PRESS

-0.62

-0.61

ascript

-0.61

aukee

-0.61

intosh

-0.59

POSITIVE LOGITS

 they

0.83

 someday

0.79

 whoever

0.77

 nobody

0.76

uces

0.73

we

0.71

soever

0.71

 everyone

0.68

 whenever

0.68

pload

0.68

Activations Density 0.043%

phrases indicating a purpose or intention

the conjunction "that" used to introduce clauses or explanations

No Comments

No Known Activations