INDEX

Explanations

phrases indicating logical reasoning or argumentation

oai_token-act-pair · gpt-3.5-turbo

expressions related to making claims and denials

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ortment

-0.84

pour

-0.69

itton

-0.67

naissance

-0.60

 largeDownload

-0.59

 newcom

-0.58

soon

-0.58

 doubtless

-0.58

nown

-0.58

Kings

-0.58

POSITIVE LOGITS

 anymore

1.64

 slightest

1.10

nor

1.07

 anywhere

1.07

any

1.04

 anything

1.00

 necessarily

0.98

 whatsoever

0.91

 anybody

0.90

 enough

0.82

Activations Density 0.209%

phrases indicating logical reasoning or argumentation

expressions related to making claims and denials

No Comments

No Known Activations