INDEX

Explanations

references to documents being obtained

oai_token-act-pair · gpt-3.5-turbo

instances of the word "obtained" in relation to documents or information

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

atomic

-0.64

ocrates

-0.64

liam

-0.64

heartedly

-0.62

mop

-0.61

 Presbyter

-0.60

 Debate

-0.60

 Mouth

-0.59

 jung

-0.58

gaard

-0.57

POSITIVE LOGITS

CLASSIFIED

0.90

racuse

0.88

tesy

0.82

>]

0.82

 obtained

0.80

 snapshots

0.76

OGR

0.75

 unfocusedRange

0.75

cffff

0.74

andom

0.72

Activations Density 0.034%

references to documents being obtained

instances of the word "obtained" in relation to documents or information

No Comments

No Known Activations