INDEX

Explanations

references to movement downwards

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "down."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

cius

-0.68

ificent

-0.66

anooga

-0.65

ista

-0.63

#$

-0.62

ament

-0.62

ature

-0.61

 Pros

-0.58

ctic

-0.58

 gadget

-0.58

POSITIVE LOGITS

stairs

1.37

graded

1.14

 stairs

1.09

grading

1.07

LOAD

1.07

river

1.00

grades

0.96

hill

0.95

pour

0.91

played

0.90

Activations Density 0.041%

references to movement downwards

occurrences of the word "down."

No Comments

No Known Activations